Since our successful official launch of SKSIS-1 in the opening morning sessions of the 2018 U of S Soils and Crops Conference, our user-base has been averaging 297 users and 560 sessions per month. 5,128 unique users within Saskatchewan have accessed SKSIS with additional users in other Canadian provinces and territories, as well as users from 60 other countries. Our user-base has been busy providing feedback and ideas for improving and expanding the SKSIS repository to incorporate new features to support the addition of existing and new value-added data (e.g. point data, salinity information, and/or land use data) that better supports agricultural land management. In many cases the availability of enhanced soil information is transformative, providing users with access to field-scale soil landscape knowledge that was previously the domain of experts.

Top-down and bottom-up predictive digital soil mapping approaches were explored for generating detailed soil maps with the potential to inform precision agricultural management in Saskatchewan.

The top-down approach involved disaggregating the legacy soil surveys using machine-learning modelling techniques and remote-sensed environmental information like elevation models, spectral imagery, and vegetative indices to predict soil classes. Disaggregation was tested at the Swift Current watershed, where both SRTM (coarse-resolution) and LiDAR (high-resolution) elevation data are available. Disaggregation models that incorporated the LiDAR elevation data produced detailed soil class maps that made conceptual pedological sense and had acceptable predictive accuracy. The disaggregation models that incorporated the SRTM elevation data were not as accurate and generated questionable looking soil maps. Soil survey disaggregation methods will be used to map soil classes in areas where high-resolution elevation data (like LiDAR and UAV-collected imagery) exists and made publicly available as a part of SKSIS-3, however, publicly available high-resolution elevation data currently only exists for small patches of the agricultural regions of Saskatchewan.
The bottom-up approach involved the development of a protocol for semi-automated predictive digital soil mapping using site-specific information. This protocol provides a framework for generating detailed soil maps of varying properties and characteristics for small-scale sites using site-specific data from those sites. It was tested at two Saskatchewan sites to predict various soil properties and achieved reasonable predictive accuracy for certain soil properties. This protocol will be further developed and be made available for public use as a part of SKSIS-3.

The Shelterbelt Decision Support System (DSS) provides users with the capability to draw a line on an SKSIS-based map to represent the placement and length of a shelterbelt. Users may design and input rows, species, etc. to plan their new shelterbelt whether it be located in the field or farmyard. The shelterbelt-planning tool illustrates farm-level scenarios for sequestering carbon based on shelterbelt species, climate, soil zones, and specific soil attributes from SKSIS. It can also input existing shelterbelt information to analyze carbon credits and mitigate Greenhouse Gas (GHG). The Shelterbelt DSS provides quick, relevant, and practical information to farmers facing common challenges in their own shelterbelt management operations such as recommendations and best practices. Although the Shelterbelt DSS is an important producer tool (e.g. protecting them from drought erosion) it also demonstrates the importance of developing shelterbelt policy as it relates to carbon sequestration and GHG emissions.

The application program interface (API) supports 3rd party communication and data exchange with the SKSIS-2 enhanced framework, enabling the integration of SKSIS soil information with user-supplied agronomic and environmental data from user groups (e.g. agricultural consultants, Government organizations, university researchers, students and producers). This data exchange capability, provided by the API, enables more complex user-data analysis and thereby improves decision making, while building on their own user-data collected. Through the new API, this value-added data can be shared to maximize producer efficiency and minimize risk.

Extension Messages:

• Continued user group feedback has led to improvements in the SKSIS user experience.
• Predictive digital soil mapping has a huge potential to inform precision agricultural applications in Saskatchewan. It is possible to create detailed maps of soil classes using high-resolution elevation models by “disaggregating” the legacy soil survey using other publicly available environmental information. The resulting soil class maps can be used to support delineation of detailed management zones within a field. Detailed soil class maps generated by disaggregating the soil survey will be made available in future phases of the SKSIS project in areas where high-resolution elevation data exists.
• A semi-automated protocol was also developed for creating detailed soil maps of various soil properties using field-specific data and machine learning. In future phases of the SKSIS project, this will be made publicly available so users can upload soil data with observed values for the soil property of interest (ex.: carbon, pH, phosphorus, etc.) along with other site data (like EM surveys or drone-collected imagery for elevation and NDVI) and the protocol will automatically train and test a range of models, output a map of the soil property based on the best performing model, and report its accuracy.
• A Shelterbelt Decision Support System is now available to help landowners understand the economic, biodiversity, and environmental benefits of shelterbelts, and thereby efficiently plan them.
• SKSIS makes trusted/qualified Saskatchewan soil data and information available to many public and private organizations/sectors, and individuals through its Application programming interface (API). To improve and expand our soil data, SKSIS depends on collaboration and contribution of data, while respecting inherited privacy and security data restrictions. To this end, the next phase of SKSIS will develop a token-based data access control system.